Complications of Hip Treatment in Children with Cerebral Palsy
Children with cerebral palsy can have multiple complications in the course of managing the hips during growth. Most of these complications are mild and delay the full recovery, but a few cause a long decrease in physical function and pain. Late diagnosis or delayed treatment is the most severe complication because it allows the hips to develop arthritis which makes the outcome of subsequent treatment less good. Other complications include the possibility of a wound infection; most heal with wound management and antibiotics with no impact on the eventual outcome. Prolonged pain after hip surgery may be due to the prominent hip plate, from arthritis as the hip joint is healing, from heterotopic bone formation, or from an undiagnosed fracture. The treatment is based on developing a good diagnosis of the pain and addressing hip pain with steroid injections if it persists for 3–4 months. Some of the children have problems eating and sleeping at night. This is best managed with nonnarcotic pain medication and amitriptyline hydrochloride (Elavil). Fractures after hip surgery may occur and are more common if the child was placed into a hip spica cast, which is seldom needed. Identifying the fracture and doing appropriate treatment will provide a good outcome. Nonunion and delayed union of the osteotomy, avascular necrosis, and heterotopic bone formation are rare complications that can usually be managed with temporary activity restriction and treating the pain.
KeywordsCerebral palsy Hip arthritis Nonunion Delayed union Femur fracture Heterotopic ossification Embolism Thrombophlebitis Hip pain Infection Leg length difference Hip stiffness
Managing the hip problems in children with CP requires careful planning and the expectation that there may be problems along the way that need to be addressed. Some of these problems are related to missed follow-up, leading to more severe hip dysplasia and degeneration. It seems clear that the more destroyed the hip is, the more difficult will be the repair or reconstruction (“Hip Reconstruction”). However, even when the follow-up is appropriate, the underlying CP remains and recurrent displacement can occur. This requires ongoing diligent monitoring and responding in an appropriate fashion and timing. In addition to these unpredictable concerns, there are also multiple areas where complications can occur ranging from fractures to infections. The goal of this chapter is to address the complications related to the management of the hip in CP. Each complication will be addressed as it might be encountered as an individual problem in a clinical management environment.
Natural History and Symptoms
Delayed treatment is most common when there is no surveillance program or the surveillance program has failed. This requires an appropriate attention to the methods and management of the surveillance program. Other causes of delayed care may be a child in whom the medical condition is very precarious and the child is not expected to survive long term; however, they slowly develop increase health. It is important when a decision is made to not intervene, that there be ongoing follow-up, so there is the ability to reconsider the decision should the clinical condition warrant. Another cause for delayed care is the fear that the child may lose physical function. This is most typically the child who is classified as having GMFCS III who maybe struggling to be functional and the family’s and therapy staff’s fear of losing the child’s functional gait ability with hip reconstruction. When the hip is having progressive subluxation in the range where intervention is indicated but the patient is not in pain and seems asymptomatic relative to the hip, it may be hard to convince the parents of the importance of proceeding with the reconstruction. By delaying, there will very likely be a less favorable outcome when the reconstruction is required. An example is this 9-year-old child whose family refused reconstruction when the outcome would clearly have been better and the rehabilitation much easier (Case 1).
Recurrent Contracture and Dislocation
The incidence of recurrent dislocation varies greatly based on the specific procedure. Failure rates with the recommended peri-ilial osteotomy, varus derotation, and soft-tissue lengthenings vary from 0% to 4% on short-term follow-up (Miller et al. 1997; McNerney et al. 2000; Jozwiak et al. 2000). Longer term follow-up of over 10 years has shown an increase rate to 12% (Oh et al. 2007). Also dislocation rate and other complications are higher in young children with complete dislocations (Dhawale et al. 2013). Recurrent dislocations are related to the severity of the acetabular deformity and the development of severe recurrent contractures over time. It is very important to continue following the children, and if they start to develop recurrent fixed contractures, these need to be addressed with repeat soft-tissue lengthenings. Also, the redislocation rate is much higher for hypotonic hips (“Hip Special Problems”), and a different level of skepticism toward the results needs to be communicated to the parents (Case 2).
If the recurrent dislocation occurs early, in the first 4 weeks postoperatively, a very careful assessment of the surgery and why the dislocation occurred is required. If the redislocation results from such severe acetabular deficiency that the reconstruction could not adequately recreate a stable acetabulum, then further attempts to provide a stable reduction should be postponed until these children are rehabilitated. If the hip then becomes painful, it would be treated with a palliative approach (“Hip Salvage Palliative”). However, if children with spastic hip dislocation had a successful reconstruction and years later developed a recurrent adduction and hip flexion contracture followed by a hip subluxation or dislocation, the treatment regimen should be quite different. In general, as these patients are followed, as soon as their maximum abduction is less than 0°, a repeat adductor lengthening should be performed. Similarly, a second bony reconstruction should be considered when the hip migration percentage becomes greater than 40%.
Recurrent, complete full dislocations are more common in children who have anterior dislocations or a hypotonic pattern of dislocation. We have treated two children with a hypotonic anterior dislocation who required a second acetabular reconstruction, and both have remained located and ambulatory over more than 5-year follow-up. Although very rare in children with CP, direct posterior dislocations also have a high rate of recurrence after the peri-ilial osteotomy. Reconstruction of a direct posterior dislocation can be accomplished using a posterior pelvic osteotomy and/or shelf arthroplasty or only bone graft, which is fixed to the acetabulum with screws. Primary and recurrent direct posterior and posteroinferior dislocations are much more common in spastic incomplete spinal cord injuries. If there is significant spasticity present in children with spinal cord injury as the etiology of the spasticity, there needs to be awareness of a much higher incidence of repeat dislocation. This higher incidence probably results from having little strength in the gluteus maximus.
Hip Wound Infections
Wound infections from hip surgery occur primarily at the site of a femoral osteotomy. Superficial wound infections also occur in groin wounds and the anterior pelvic osteotomy site; however, these infections readily respond to local wound care, cleaning, and keeping the wound dry. Often, these are suture infections that will resolve once the suture is removed. Prophylactic antibiotics, usually a broad-spectrum cephalosporin, should be used for one dose immediately preoperatively followed by two doses postoperatively. Prophylactic antibiotics have been demonstrated to decrease the risk of infection in hip surgery in children with CP (Beauchesne et al. 1992).
Deep wound infections in the adductor wounds are rare, occurring in approximately 1 in 400 open adductor tenotomies, and almost all of these are in revision adductor tenotomies. We have not seen a deep wound infection in a primary open adductor tenotomy in children with CP. Using a careful closure of a longitudinal incision in the fascia and a good, careful, tight closure of the transverse subcutaneous tissue and a good subcuticular skin closure are important in avoiding infections in the adductor wound. It is important to prevent the deep hematoma from leaking out into the wound and becoming secondarily infected. It is also crucial to keep a sterile, sealed dressing in place using an adhesive-backed plastic for the first 5–7 days postoperatively so the adductor wound does not become soiled with stool or urine. The main reason deep wound infections occur in revision surgery is that there is often a great deal of scar with no good fascial layer, which makes good, tight closure at the conclusion of the procedure much more difficult. There is often much more aggressive release required as well, which puts more tension on the closed skin wound.
Adductor Wound Infections
If a deep wound infection occurs with gross purulent drainage from the deep muscle lengthening site, the children should be returned to the operating room and the wound opened widely, debrided, and irrigated. If there is significant necrotic tissue that cannot be debrided, the wound should be packed with an iodine-soaked solution, and then these children must be returned again for a dressing change in 48 h. When the wound has good, clean tissue, the skin and subcutaneous tissue can be closed tightly in one layer with nonabsorbable sutures. The deep wound dead space should be closed using a suction drain, which is brought out distally in the medial thigh. These drains should be attached to constant suction; we have not used any irrigation. Usually, the drainage decreases significantly after 24–48 h; however, the drain should be left in place until there is almost no drainage at all. This often requires 3–5 days, but the drain should not be left in longer than 7 days because of the risk of secondary infection. Throughout this process, children should be treated with intravenous antibiotics based on the culture results from the wound. Antibiotics are used for approximately 7 days after the drain is removed. Another alternative to the wound closed over a drain if there is excessive and a large amount of purulence is to use a wound VAC dressing. Although we have never seen a deep wound infection from a pelvic osteotomy, if it should occur, the same approach for treating that wound is recommended.
Femoral Osteotomy Infections
Deep wound infections at the site of the femoral osteotomy are much more common, occurring in approximately 2% of femoral osteotomies at our facility (Beauchesne et al. 1992; Miller et al. 1997). Most of these infections occur in nonambulatory children who are very thin. When it is clear that there is a deep infection, the wound should be opened widely and treated with dressing changes, usually starting with an iodine-soaked dressing or a VAC dressing, and antibiotics. If children present with only local symptoms with a wound that is opening and freely draining with minimal cellulitis and no systemic fever, the wound can be opened in the outpatient department and dressing changes started along with oral antibiotics. If children are febrile, or have a significant local cellulitis, they should be admitted to the hospital and started on intravenous antibiotics with wound opening and packing. As the wound cellulitis resolves and the children become afebrile, they can be discharged home on the appropriate oral antibiotics as determined by the result of the wound culture. As the purulent drainage decreases, dressing changes should be switched to saline to allow the development of healthy granulation tissue. After 7–10 days, when the wound no longer has any cellulitis, the antibiotic may be discontinued. No attempt should be made to close this wound back over the plate, nor should the plate be removed until the osteotomy has healed. Sometimes the wound will close over on its own, but in our experience, there is a high rate of recurrent infection so long as the plate is in place.
When radiographs show adequate healing of the osteotomy site, children are brought back to the operating room and all the hardware is removed. The wound can be loosely closed, and the children are again given oral antibiotics based on the results of the culture at the time the plate is removed. Sometimes there is a significant amount of drainage and some necrotic bone, all of which can be well irrigated and cleaned out at the time of the plate removal. This drainage does not need to be treated as a deep osteomyelitis, and in every child whom we have treated following hardware removal, the wound has closed within 2–3 weeks. We have never seen a recurrent infection. After the wound has closed completely, usually in 2–3 weeks, the oral antibiotics are discontinued. We have never seen a child in whom the osteotomy would not heal, even if the wound was left open with an exposed plate. However, these children are often uncomfortable while the plate is exposed, especially with range of motion and ambulation. It is important to continue maintaining and gaining range of motion and pushing the children into ambulation, standing, and walking as much as they will tolerate. This movement helps with the healing process of the bone.
Femoral Osteotomy Nonunions
Nonunions of the femoral osteotomy, using the described technique (Beauchesne et al. 1992), occur in approximately 1 in 300 osteotomies based on our experience. Approximately the same number of recognized delayed unions occur. There is no definite pattern of occurrence; however, there are several important factors that will help avoid nonunions. First, it is important to use a large enough plate so that it will not fail by breaking or pulling free of the bone before the union occurs. Good compression of the medial cortex at the time of the internal fixation must be ensured, and there should always be at least good opposition of the medial cortex with slight medialization of the distal fragment. Importantly, the distal fragment should not be allowed to lateralize because this provides a very poor mechanical construct (Case 3). Delayed unions, when they occur, usually require approximately 6 months to go to union. There is no definite time for determining that a delayed union has become a nonunion; however, a good rule is that children should have asymptomatic union by 6 months postoperatively (Case 3).
Assuming that there should be an asymptomatic union of the bone by 6 months postoperatively, a cutoff point was arbitrarily chosen to make the diagnosis of a nonunion. This cutoff point is any child whose femoral osteotomy site has continued evidence of nonhealing on radiographs and is symptomatic. If nonunion occurs, the children are returned to the operating room where the plate is removed and a larger or more stable plate inserted with strong compression, followed by bone grafting along the anteromedial aspect of the osteotomy site. Taking down a fibrous nonunion is not recommended, and the nonunions that we treated have healed well with just repeat compression and applying anteromedial bone grafting. We used crushed bank bone; however, the use of iliac crest bone may provide better osteogenic potential. The important aspect of this treatment is the proper rigid internal fixation with compression and returning the children to weightbearing status. Some parents will be hesitant to push their children back into standing because they may believe some of the problem was caused by weight bearing. However, most of the delayed unions and nonunions occur because of technical errors at the time the osteotomy was performed and not because of the children’s activity.
There is increasing use of locking screw – plate technology, some incorporating blade plates (Zhou et al. 2015) and others using only screws and side plates (Haefeli et al. 2010). When these systems are compared directly, traditional plate blade and the locking screw and side plate seem to have similar outcomes (Rutz and Brunner 2010). Based on antidotal evidence, of more people seeming to report delayed and nonunions with locking plate technology, further larger series are required to be sure the rate of healing is the same since it is hard to get good compression with the locking plates when the distal fragment is medialized.
Fractures of the Femur
Fractures of the proximal femur that occur as a complication of femoral osteotomy are most common surrounding the blade plates. These fractures may be acute fractures or bone failures, or fractures occurring from stress risers with the plate in place and a healed osteotomy. Occasionally, a fracture may also occur after plate removal.
Acute fractures are defined as occurring before the osteotomy union is solid. In our experience, these fractures occur in approximately 1 in 200 osteotomies and, in almost every case, are due to technical errors at the time of the osteotomy.
If the osteotomy is made too close to the insertion of the blade plate, there may be a fracture of the lateral bridge of the proximal fragment (Case 4). This fracture can be avoided by ensuring that there is an adequate lateral bridge or ensuring that the distance from the insertion of the blade plate to the osteotomy site is wide enough. A common error is measuring the hypotenuse of this triangle instead of its direct right angle leg, thereby getting a lateral bridge that is too narrow (“Hip Surgical Atlas”). The fixation of the blade plate requires a strong lateral cortical buttress against which compression can be applied. If a lateral fracture occurs and there is room in the femur, the blade plate should be moved more proximally and placed into the center of the femoral neck as far as possible into the femoral head. Often, this means that a plate with more added valgus has to be used and a lateral cerclage wire is generally required because there is not enough lateral bone for the compression. It is important for surgeons to remember that the blade plates may be bent into more of a valgus or varus position if the need arises. The problem with putting too much valgus into the plates, however, is that the ability to compress the osteotomy site, which is important to prevent nonunions, is lost. This is why these plates are made only in 90° and 100° angles.
Fracture of the greater trochanter occurs when the insertion point of the blade is into or too close to the apophysis of the greater trochanter. The fracture may then propagate into the femoral head, with a fragment of the femoral head and neck and trochanter elevating, or only the greater trochanter may fracture free of the plate (Case 5). This fracture can be avoided by never inserting the blade into the apophysis of the greater trochanter and by always staying in or below the subchondral bone of the greater trochanteric apophysis. If the fracture goes into the greater trochanter only, and the greater trochanter lifts off significantly, it can be repaired by an open reduction using cerclage wires or cables to bring the trochanter back down to the plate. However, if the fracture propagates into the femoral neck and head, such that the proximal fragment includes a component of the femoral neck and head with the greater trochanter, an open reduction with exchange of the plate or using a locking screw plate is required. In this open reduction and plate exchange, screws are placed into the proximal fragment and a new plate is inserted, usually in a much more valgus position, into the femoral head along with lateral cerclage wires. This open reduction is somewhat complicated to perform, and it is often helpful to have the middle piece between the proximal fracture and the distal osteotomy fixed to the distal fragment using a small anterior plate. This fracture should be diligently avoided by ensuring that entering too far proximally into the apophysis of the greater trochanter does not occur.
Distal End of Plate Fractures
Later fractures that occur as stress risers from the plate site with the blade plate in place should be treated with removal of the blade plate and repeat open reduction with a device most appropriate for the fracture pattern. Likewise, fractures immediately after plate removal should be treated as de nova fractures and usually require an open reduction or internal fixation with repeat plating, or occasionally the use of intermedullary flexible nails. It is crucial to ensure that the proximal fragment is not allowed to fall into too much varus or lose the derotation that had been obtained.
Leg Length Discrepancy
Leg length discrepancy secondary to problems of the hip may be due to asymmetric contracture of hip muscles, asymmetric varus osteotomy, hip subluxation or dislocation, or suprapelvic pelvic obliquity.
Caused by Adductor Contracture
A very common cause of perceived leg length discrepancy occurs secondary to asymmetric contracture of the hip abductors or adductors. This contracture leads to sitting or standing positions in which a great limb length discrepancy is apparent. When this contracture is mild, it is often better to not accommodate the contracture in seating with asymmetric wedges or in standing with a shoe lift. As these contractures and apparent limb length discrepancies get more severe, some accommodation with mild shoe lifts may be needed. In general, no more than half of the apparent limb length discrepancy should be corrected, because if too much correction is made more rapid development of more asymmetry is fostered. Likewise, in seating, if children develop an apparent limb length discrepancy, some of the adduction and contralateral abduction to keep them sitting in a fairly neutral position should be accommodated. If an attempt is made to keep the legs in a perceived normal position, the pelvis and trunk will often rotate, causing the abducted side to rotate forward so that these children are sitting in a sideways position in the wheelchair. As these asymmetries become more functionally significant, they need to be addressed with soft-tissue releases.
Secondary to Varus Osteotomy
The best treatment of limb length discrepancy occurring secondary to a unilateral varus osteotomy is avoiding doing unilateral osteotomies. Almost all children should have symmetric surgery; however, a derotational osteotomy, in which some inadvertent varus was obtained, may occasionally be performed unilaterally. This inadvertent varus is typically less than 1 cm and should not cause any functional difficulty. Sometimes, however, this slight limb length discrepancy from the varus osteotomy may be magnified by asymmetric adduction contracture and will need to be accommodated. If a major asymmetric varus osteotomy was performed, there may be as much as 3–4 cm of limb length discrepancy. This degree of limb length discrepancy has to be addressed for standing with the use of a major shoe lift and in sitting with a cutout of the seat. If this discrepancy occurs in young growing children, the leg length discrepancy is equalized by contralateral femoral epiphysiodesis after careful monitoring with scanograms. However, it is difficult to predict precisely how much remaining growth is present in nonambulatory children. The use of temporary plate epiphysis is another option which can be used, and it is helpful to determine when the correct leg length difference has been reached. Then the plate can be removed and growth will return. Another option to gain leg length equality is a varus shortening osteotomy on the long side.
Leg length discrepancy may be a sign of a dislocated or subluxated hip, which has to be ruled out with an appropriate radiograph. When hip dislocation is causing limb length discrepancy, the hip needs to be treated according to the indications for treatment previously discussed.
Leg length discrepancy secondary to pelvic obliquity may be caused by asymmetric contractures in the windblown deformity. However, this discrepancy also occurs as a suprapelvic pelvic obliquity coming from significant scoliosis. In general, children with CP who develop a suprapelvic pelvic obliquity actually tend to lean into the scoliosis in such a way that the pelvis may be relatively straight when they are sitting. Seat cutouts may often be required. Other adaptive mechanisms to accommodate this pelvic obliquity may also be required until the discrepancy is surgically corrected. This deformity is especially difficult in those children who have had a spine fusion in which the pelvic obliquity or rotation was not corrected.
Heterotopic ossification in children with CP has been a problem primarily at the hip. It tends to occur after hip surgery, especially if the hip surgery is done concurrently with or in close proximity to spinal surgery. Ossification of the hip has been reported to occur following hip surgery that is concurrent with spinal fusion (Krum and Miller 1993) or dorsal rhizotomy (Payne and DeLuca 1993). In rare situations, heterotopic ossification may develop in the hips after spinal fusion with no concurrent hip surgery (Krum and Miller 1993). This ossification occurred in two children whom we have seen, and it tends to lead to severe heterotopic ossification and complete fusion of the hip joint and has been reported after adductor lengthening (Ushmann and Bennett 1999).
After Adductor Lengthening
Heterotopic ossification after adductor, iliopsoas, and proximal hamstring lengthening is extremely common. The most common source of this heterotopic ossification is along the tendon sheath of the iliopsoas. This ossification is rarely a clinical problem; however, there may be some prolonged discomfort for 3–4 months as the heterotopic ossification matures. Some children will have pain longer during active range of motion, especially with forced hip flexion. A very long, thin piece of heterotopic bone may develop in the sheath of the iliopsoas in some of these children, and we have seen several cases in which a fracture of this long piece of heterotopic ossification developed. When this fracture develops, it often causes pain or discomfort for approximately 3 or 4 weeks and then resolves. The heterotopic ossification of the iliopsoas rarely requires any supportive or interventional treatment beyond using occasional acetaminophen or ibuprofen for pain control and continuing with gentle range of motion. Heterotopic ossification in the iliopsoas tendon sheath can be decreased by ensuring that the tenotomy is performed well away from the apophysis of the lesser trochanter.
When heterotopic ossification occurs at the site of the proximal hamstring lengthening, it is usually much more severe and somewhat more difficult to treat, but fortunately it is also much more rare (Lee et al. 1992). These children’s hips are often very painful for no apparent reason, but a workup including a bone scan will show the development of substantial heterotopic ossification. The typical scenario is children who, at 4–6 weeks postoperatively, have a normal radiograph but are continuing to have severe pain at the hip with any activity. A bone scan, which should often be obtained at this time, may confirm the heterotopic ossification by showing very hot uptake in the area of the surgery site (Case 7). At the time when the bone scan is hot but the radiograph is normal, there is no benefit from the use of bisphosphonates or radiation because the process is already too far along. These children instead should be started on maximum antiinflammatories, usually using ibuprofen or naproxen. Although indomethacin may be better for treating heterotopic ossification, it is not approved for use in children and is not sufficiently better than approved drugs.
Often, the discomfort will make sleeping and eating difficult. During the most acutely painful phase, a narcotic analgesic such as acetaminophen with codeine or oxycodone may be needed. An additional moderate to high dose of diazepam is required to decrease any signs of spasticity so that the muscles will stay relaxed and will not further irritate the heterotopic ossification. After 8–10 weeks, this discomfort should start to diminish and the medication, especially the narcotics and diazepam, should be weaned. If children are still having problems at this point, an antidepressant, usually amitriptyline hydrochloride (Elavil), should be started. Amitriptyline is an excellent drug to promote pain control and improve sleep and general attitude. During this period, gentle range of motion should be performed as much as possible to avoid the development of a fused hip. Gradually, as the active process decreases and if the hip has not gone on to full fusion, the hip range of motion should start increasing (Case 7).
Other common situations in which heterotopic ossification may occur are associated with proximal femoral resections (Widmann et al. 1999; McCarthy et al. 1988; Perlmutter et al. 1993; Nakajo and Endo 1969), adductor lengthenings (Ushmann and Bennett 1999), and exuberant callus at a femoral osteotomy site (Payne and DeLuca 1993) and along the capsulotomy site following femoral reduction. The gluteal fossa of the ilium may also develop heterotopic ossification after perforation of the ilium with the pelvic end of a spinal rod.
Prophylactic Treatment of Heterotopic Ossification
One specific recommendation for the prevention of heterotopic ossification is avoiding concomitant hip surgery with spinal surgery (Payne and DeLuca 1993; Krum and Miller 1993). Either a release of some factors from the spinal surgery area or the magnitude of the procedure increases the incidence of hip ossification. Removing lateral blade plates at the time of spinal surgery is appropriate; however, no other hip surgery should be performed for at least 4 months following spinal fusion or dorsal rhizotomy. After 4 months, the risk of heterotopic ossification diminishes. We have not seen any increased risk of heterotopic ossification in hip surgery if it is performed 4 months or later following spinal surgery.
It is not clear if there are children who might benefit from other prophylactic treatments of heterotopic ossification. Children who have developed heterotopic ossification from muscle surgery and are now required to have more substantial hip surgery, such as femoral resection, are at very high risk for developing significant heterotopic ossification. For these children, prophylactic radiation treatment on postoperative day 1 or 2 may be considered. It is difficult to identify exactly which of these children will develop significant symptomatic heterotopic ossification. We have no experience using bisphosphonates, and, based on published data, we are not very optimistic that they would be useful (Banovac 2000; Thomas and Amstutz 1987).
Postoperative Hip Pain
Hip pain is present in all children after hip surgery, and control of this pain is a mandatory part of the orthopedic management of these children. The standard pain treatment program should anticipate that it will take 6–8 weeks after surgery until most of the pain is resolved. If there continues to be a significant amount of pain present by 8–12 weeks after hip reconstruction or muscle lengthening, the cause of this pain needs to be specifically diagnosed and treatment designed based on the diagnosis. Many potential causes of this pain can be identified.
The development of heterotopic ossification should be suspected, especially if children are continuing to have severe pain after only having muscle surgery. If radiographs are normal and heterotopic ossification is suspected, a bone scan, which will identify the early stages of heterotopic ossification, should be obtained (Case 7).
Medial Plate Protrusion
Medial protrusion through the calcar or the femoral neck by the blade plate may cause pain by producing an iliopsoas bursitis. This bursitis is most typically a problem in children who have had derotation to improve their walking ability but continue to have increased pain 3–9 months after surgery and are not quite making the rehabilitation progress expected. These children typically refuse to stand with the hip fully extended. Often, the primary complaint is not pain but rather the inability to make progress in rehabilitation, especially in the ability to gain straight upright standing. On physical examination, it is often very difficult to localize the problem because when these children are relaxed they have full hip range of motion with no pain. There usually seems to be no reason why these children should not be able to make progress in rehabilitation or gain straight upright stance. Radiographs may show only some slight medial protrusion of the blade plate, and, occasionally, if the rotation of the femur is not correct, this may not even show as the plate is directed slightly anteriorly. In this clinical scenario, the plate should be removed once the osteotomy has healed, and in some children, dramatic progress in rehabilitation is then obtained (Case 8).
Acute degenerative arthritis may occur and cause severe hip pain in the rehabilitation phase between 3 and 6 months following hip reconstruction. Typically, these children do as expected, gaining range of motion and improved comfort until approximately 6–10 weeks postoperatively, when the hip pain slowly starts getting worse. By approximately 4 months after surgery, the pain may be so severe that any movement of the hip joint causes pain. In some of these children, a small range of motion is comfortable, but as soon as the hips are moved outside this window, they are very painful. Radiographs will typically demonstrate some narrowing of the hip joint space. This narrowing often occurs in children who have a small ridge identified at the level of the triradiate cartilage in the acetabulum.
During reconstruction, good coverage and reduction of the femur is obtained, but the femoral head sits somewhat laterally against this medial ridge. As the range of motion is started, this medial ridge is worn down, causing symptoms of degenerative arthritis and synovitis in the hip joint. When these symptoms are identified at the initial stage, antiinflammatory medication should be started following an antiinflammatory dose schedule usually using ibuprofen or naproxen. The hip joint should be injected with a deposteroid or 80 mg triamcinolone acetate, with a small dose of approximately 1 ml bupivacaine hydrochloride. This local anesthetic injection will quickly demonstrate that acute degenerative arthritis is the source of the pain, as the pain should be gone for 6–8 h. A significant decrease in the pain should be expected in 48–72 h after steroid injection.
The hip joint injection of the steroids and bupivacaine hydrochloride can be performed in the outpatient clinic if physicians are confident that they can palpate the anatomy of the hip joint and are able to enter the hip joint. However, in older children or in children with less-clear landmarks, it is better to perform the injection in the radiography suite under fluoroscopic control. Steroids can be injected every 4 weeks for up to three injections if the pain has not made substantial improvement. At the same time, if the children are also having trouble sleeping and are eating poorly, an antidepressant, typically amitriptyline hydrochloride (Elavil) twice a day, should be started. The antidepressant will improve pain control, sleep, and general attitude.
The outcome of treatment in this scenario has a very high success rate, with complete resolution of the hip pain in 3–6 months. Substantial remodeling of the hip joint with recreation of hip joint space often occurs as new cartilage seems to heal in the hip joint. However, this remodeling really only works in children who have open growth plates, and we would be very hesitant to expect this kind of outcome in adults. We have had no experience using this regimen except in children with open growth plates. At 1 year after reconstruction, in spite of these problems, there is usually good recreation or maintenance of hip joint space on radiographs (Case 9).
Sudden Pain in Therapy
Continued or increased pain in the hip following hip reconstruction in children who are noncommunicative can be a real challenge. Caretakers often will feel that the pain is coming from the hip, but a full examination does not seem to demonstrate any evidence of pain. When pain exists in noncommunicative children, other sources of pain need to be considered and the appropriate and typical workup should then proceed.
Avascular necrosis has been reported following spastic hip reconstruction (McNerney et al. 2000; Hesketh et al. 2016a); however, we have also seen a definite case of avascular necrosis (Case 10 AVN). Determining risk factors and a clear diagnosis is difficult. Most reports of avascular necrosis used cast immobilization, which may explain part of its cause. Abducting spastic hips against the spasticity has been shown to produce avascular necrosis in insensate children (Rink and Miller 1990). If avascular necrosis develops, the recommended treatment is to continue with range of motion and wait for the avascular necrosis to run its course and then address the remaining residual problems (Case 11). Sometime hip resection is required for a severe case (Case 10).
Lateral femoral head collapse may also occur when there is an insufficient capsulotomy and postoperative rehabilitation attempts to stretch the contracted medial capsule. Often, the lateral aspect of the severely subluxated femur is very osteoporotic, and when high pressure is applied to this osteoporotic femoral head, it may be caused to collapse. This collapse can be prevented with an adequate medial capsulotomy to allow adequate abduction. We initially presumed that 0° of abduction was sufficient intraoperatively to not require a capsulotomy; however, we now believe that abduction to at least 20° or 30° in the operating room is possible. This means more medial and more aggressive capsulotomies need to be performed, which has not led to avascular necrosis and has decreased the incidence of finding compression of these osteoporotic lateral femoral heads. It is impossible to ensure that some collapse of the lateral femoral head was not due to avascular necrosis; however, this collapse does not have the appearance of a full avascular necrosis as reported (McNerney et al. 2000). This may also explain why avascular necrosis occurs when these osteoporotic femoral heads are forced into the acetabulum and then held in a fixed cast, which may limit the blood flow to the femoral head and why some reports show a very high incidence (Koch et al. 2015). When lateral collapse occurs, it usually creates a flattening of the superolateral border of the femur, limiting abduction similar to the protrusion that becomes present in the hip of Perthes disease with lateral extrusion. If this collapse occurs and abduction becomes limited, the best treatment is to do a valgus osteotomy to accommodate the deformity. Separating true ANV from osteoporotic collapse is very difficult (Hesketh et al. 2016a).
Thrombophlebitis and Pulmonary Embolism
It is rare to see children or young adults up to age 21 years with CP who have had thrombophlebitis or pulmonary embolism. When it occurs, a full coagulation work is required as all the children we have seen with thrombosis had a further diagnosis of a congenital hypercoagulable condition. There are also no reports in the literature that substantiate this occurrence. We have had two children with severe heterotopic ossification who were initially diagnosed as having deep venous thrombosis based on an ultrasound Doppler flow study. A bone scan should be obtained on any child with swelling in the legs, in whom deep venous thrombosis is believed to be present, to rule out heterotopic ossification before we would consider anticoagulation therapy. The consideration for deep venous thrombosis treatment is contraindicated until heterotopic ossification has been ruled out by a bone scan or a clear thrombosis is visualized. There is one literature of a young adult with thrombosis after femoral resection; further workup was not defined (Knaus and Terjesen 2009).
Hip Joint Stiffness
Persistent decreased range of motion after hip reconstruction may be due to heterotopic ossification. Otherwise, this decreased range of motion may also be caused by pain and acute degenerative arthritis that was previously discussed. Children who are placed in prolonged cast immobilization may develop stiffness of the hip that cannot be mobilized. With the elimination of casting, the problem of hip stiffness and limited range of motion almost completely disappears. There are several other procedures, especially the shelf arthroplasty and the Chiari osteotomy, that have a very high risk for developing a severe limitation of hip range of motion. Therefore, these procedures should be avoided as well.
Case 1 Suzi
Case 2 Wade
Case 3 Jose
Case 4 Darris
Case 5 Shauna
Case 6 Samuel
Case 7 Nathan
Case 8 Gabriel
Case 9 Aaron
Case 11 Jeffrey
- Hesketh K, Sankar W, Joseph B, Narayanan U, Mulpuri K (2016b) Inter-observer and intra-observer reliability in the radiographic diagnosis of avascular necrosis of the femoral head following reconstructive hip surgery in children with cerebral palsy. J Child Orthop 10:143CrossRefPubMedPubMedCentralGoogle Scholar
- Thomas BJ, Amstutz HC (1987) Prevention of heterotopic bone formation: clinical experience with diphosphonates. Hip 1987:59–69Google Scholar